Systematic Literature Review in the Development of Datasets and Fact Verification Models for Indonesian Language

This study explores the major challenges in fact verification for the Indonesian language due to gaps in managing dynamic information, uncertainties from linguistic variations, and the proliferation of online narratives. Most text classification research for fact verification relies on English-language datasets, which may not be suitable for the Indonesian context. Through a Systematic Literature Review (SLR) approach, this research analyzes the development of datasets and models for Indonesian-language fact verification tasks. Utilizing the PICOC framework, this research conducted a systematic literature review to explore the development of datasets and models for Indonesian-language fact verification tasks. The main objectives of this research are to identify existing research shortcomings, evaluate the performance of current fact verification models, and propose recommendations for future research. These findings highlight the lack of research specifically addressing fact verification in Indonesia, as well as the lack of datasets and models specifically designed for the fact verification context. Existing models often perform poorly in terms of precision, recall, and F1 score, with average scores below 70%, indicating difficulty in identifying relevant evidence and consistently verifying facts. The key findings indicate a lack of research specifically addressing fact verification in Indonesia, as well as limited datasets and models specifically designed for fact verification. Additionally, most fact-verification tasks do not cover specific topics, except those related to politics. Common performance metrics used in this analysis include accuracy, precision, recall, F1-Score, and Dataset Score.